Principle Technologies responded to the benchmarking controversy surrounding its recent gaming benchmarking report. The report, which was paid for by Intel, pitted AMD's Ryzen processors against the Intel's new lineup of Coffee Lake Refresh processors. A closer examination of the results revealed a few test conditions that obviously could skew the results in favor of Intel's processor, including using a less-capable CPU cooler on AMD's chip, disabling half the cores on an AMD Ryzen processor, and a listing in the test notes that said the firm overclocked the RAM on Intel's platform only. The company says that it is retesting the processors to correct the errors.
The firm responded to all three allegations, but the most important revelation revolves around AMD's Game Mode. This feature essentially disables half of the processors' available cores to circumvent the eccentricities of the Threadripper architecture, but isn't designed for use with the mainstream Ryzen processors. Principled Technologies admitted that it did test the Ryzen 7 2700X in Game Mode, which turns the eight-core processor into a four-core chip, thus crippling it in gaming tests that respond well to extra cores and threads. The company is working to retest the Ryzen processors with the chip in its native eight-core mode.
The company also explained its decision to use a less-capable cooler with the Ryzen 7 2700X. As we know, less-capable coolers can impact performance, but the company stuck with the stock AMD cooler because "AMD said it was a good cooler." That will certainly need to be rectified during the company's retests, as the beefier Noctua cooler on the Intel processor could afford it an advantage. As we've proven in the past, improved cooling benefits both AMD and Intel's chips by allowing the processors to operate at their Boost frequencies more frequently, and then maintain the heightened clock speeds for longer periods of time.
The company also clarified the memory settings. Principled Technologies used the XMP profile, which automatically assigns tight timings for the Corsair Vengeance DDR4-3200 memory (running at DDR4-2666) installed in the Intel system. However, its initial report did not disclose that the company actually used the D.O.C.P. settings for AMD's XMP-equivalent, meaning that the memory tests were in fact fair.
The firm also clarified the motherboards it used for testing the AMD systems, which were initially listed incorrectly. We'll follow up as more information becomes available.
The Noctua cooler on the AMD wouldn't have made any significant difference in the results.
I agree. AMD needs to rework the software to recognize the CPU and either not allow it to be turned on for mainstream CPUs or maybe just turn off SMT, although I doubt that would be beneficial unless it allows them to clock higher.
I think they should use the same fan as I don't know anyone who plans to buy a 9900K and use the stock Intel fan. Most go for the high end Noctua or some AiO.
That said I don't think it would give either an advantage as the turbo speeds are programmed and the stock coolers are set to allow them their distinct turbo speeds at certain settings. The only benefit I can see is if they overclocked the system.
Hope they can make it right this time, but still suspicious...
But, I would say the use of the stock AMD Wraith Prism cooler (a somewhat beefy cooler from AMD, they call it their own "ultimate cooling solution") is nothing to blame Principle Technologies for purposefully fudging the results. AMD says that better cooling allows automatic OVERCLOCKING. It seems their only mess up was with AMD's own "Game Mode".
The Wraith cooler is fine and sufficient for normal use. But the Noctua cooler is on a higher level. The coolers aren't equivalent and thus the Intel CPU received more cooling headroom for its boost tech to take advantage of.
Cooler quality does have an impact on performance with both AMD and Intel systems. Turbo Boost, Precision Boost, and XFR all respond to the thermal conditions of the chip, among other factors, to boost. Better coolers can result in more frequent boosting, higher boost, and longer boost duration.
As for why they even bother including it in Ryzen systems. Im pretty sure all hyperthreaded computers have the option for threading to be turned off including Intel's. AMD's way of turning it off in gaming mode just seems to be bad design (at least for non-threadripper CPUs).
I'm pretty sure I remember clicking on game mode without threadripper that it asks you to also enable "legacy compatibility mode." I'll need to double check that when I get home.
But regardless, PT are supposed to be professionals and they already worked with AMD to review Threadripper and enabled Game Mode accordingly. They can't claim ignorance to disabling half the cores because they should know what Game Mode does from testing Threadripper. They should have done just one test with game mode on and off to see what the difference would be.